[pacman-dev] New Optional Dependencies Aren't Logged.
`pacman -Su' suggested an optional dependency of gst-plugins-bad. I looked to see what that was and didn't want it. Later, due to discussion on the local LUG IRC, I wanted to look up what package had the new optional depends but found /var/log/pacman.log didn't record it. Please consider logging them. Looking at the optional depends of all the upgraded packages, I think it was webkit2gtk, but if it were logged then I could check if there were any new optional depends in the more general case, e.g. looking back in time. -- Cheers, Ralph.
On 03/17/19 at 03:31pm, Ralph Corderoy wrote:
`pacman -Su' suggested an optional dependency of gst-plugins-bad. I looked to see what that was and didn't want it. Later, due to discussion on the local LUG IRC, I wanted to look up what package had the new optional depends but found /var/log/pacman.log didn't record it. Please consider logging them.
Looking at the optional depends of all the upgraded packages, I think it was webkit2gtk, but if it were logged then I could check if there were any new optional depends in the more general case, e.g. looking back in time.
Based on your description, this sounds like a case of "I was mildly curious about something this one time, so please make a permanent record of this information on every user's system in case I happen to be curious about it again in the future." I understand the desire, but there are already tools to find what a package optionally depends on and what optionally depends on a package. Before we start bloating every user's log with extraneous information, there needs to be a more substantial use case than just idle curiosity.
Hi Andrew,
Based on your description, this sounds like a case of "I was mildly curious about something this one time
No, it would have been useful to know. There's no need to overly weaken it.
I understand the desire, but there are already tools to find what a package optionally depends on and what optionally depends on a package.
Have you one in mind? It's not pactree(1), I checked that first.
Before we start bloating every user's log with extraneous information
Point taken, though looking at pacman.log, it would be very slight bloat given all its existing content. -- Cheers, Ralph.
On 3/17/19 12:49 PM, Ralph Corderoy wrote:
Hi Andrew,
Based on your description, this sounds like a case of "I was mildly curious about something this one time
No, it would have been useful to know. There's no need to overly weaken it.
I understand the desire, but there are already tools to find what a package optionally depends on and what optionally depends on a package.
Have you one in mind? It's not pactree(1), I checked that first.
Before we start bloating every user's log with extraneous information
Point taken, though looking at pacman.log, it would be very slight bloat given all its existing content.
What you want is an advanced metadata crawler/formatter for the pacman database. Simply logging it wouldn't help in the common case, as you'd still need to compare the log info with the current state of the system and figure out what is relevant. Fortunately, there is just such a tool, which has only one problem -- that being, that people aren't aware of it. :p Try out the pacutils package, specifically the paccheck program: paccheck --recursive --quiet --opt-depends $pkg It will list optdepends that are not fulfilled, recursively, for the package(s) you specify, and has a lot of other helpful options as well. -- Eli Schwartz Bug Wrangler and Trusted User
Hi Eli,
Fortunately, there is just such a tool, which has only one problem -- that being, that people aren't aware of it. :p Try out the pacutils package, specifically the paccheck program:
paccheck --recursive --quiet --opt-depends $pkg
Thanks, now installed. I'll look over all the package's commands. $ paccheck --quiet --opt-depends | fgrep gst-plugins-bad webkit2gtk: unsatisfied optional dependency 'gst-plugins-bad: media decoding' $ (A shame it doesn't have POSIX options like `-q' for the common ones.) -- Cheers, Ralph.
On 03/17/19 at 04:49pm, Ralph Corderoy wrote:
Hi Andrew,
Based on your description, this sounds like a case of "I was mildly curious about something this one time
No, it would have been useful to know. There's no need to overly weaken it.
I'm not trying to weaken anything; you simply haven't told us why you needed the information beyond "I wanted to". If there was an actual problem you were trying to solve, describing it might help us come up with a better solution.
I understand the desire, but there are already tools to find what a package optionally depends on and what optionally depends on a package.
Have you one in mind? It's not pactree(1), I checked that first.
Depends somewhat on exactly what you need. Based on the information you have provided, it sounds like -Qi/-Sii would have been enough, but paccheck would be another potentially useful tool.
Before we start bloating every user's log with extraneous information
Point taken, though looking at pacman.log, it would be very slight bloat given all its existing content.
Hi Andrew,
I'm not trying to weaken anything; you simply haven't told us why you needed the information beyond "I wanted to". If there was an actual problem you were trying to solve, describing it might help us come up with a better solution.
Okay, sorry.
it sounds like -Qi/-Sii would have been enough
Oh, the second -i adds the `Optional For' output. pacman(1) omits that useful fact. Does this count as a report to get it added? -i, --info Display information on a given package. The -p option can be used if querying a package file instead of the local database. Passing two --info or -i flags will also display the list of backup files and their modification states. So I could have done `pacman -Sii gst-plugins-bad' and then looked at the `Optional For' to see what of those I had installed and were upgraded recently. Thanks. -- Cheers, Ralph.
On 03/17/19 at 06:02pm, Ralph Corderoy wrote:
Hi Andrew,
I'm not trying to weaken anything; you simply haven't told us why you needed the information beyond "I wanted to". If there was an actual problem you were trying to solve, describing it might help us come up with a better solution.
Okay, sorry.
it sounds like -Qi/-Sii would have been enough
Oh, the second -i adds the `Optional For' output. pacman(1) omits that useful fact. Does this count as a report to get it added?
-i, --info Display information on a given package. The -p option can be used if querying a package file instead of the local database. Passing two --info or -i flags will also display the list of backup files and their modification states.
You're looking at the -Q options; listing reverse dependencies with -ii only applies to -S and is documented in that section.
So I could have done `pacman -Sii gst-plugins-bad' and then looked at the `Optional For' to see what of those I had installed and were upgraded recently. Thanks.
Hi Andrew,
Oh, the second -i adds the `Optional For' output. pacman(1) omits that useful fact. Does this count as a report to get it added?
-i, --info Display information on a given package. The -p option can be used if querying a package file instead of the local database. Passing two --info or -i flags will also display the list of backup files and their modification states.
You're looking at the -Q options
You're right; not the first time I've got that wrong.
listing reverse dependencies with -ii only applies to -S and is documented in that section.
Okay, but I still don't think -Sii is accurately described in pacman(1). :-) It has SYNC OPTIONS (APPLY TO -S) ... -i, --info Display information on a given sync database package. Passing two --info or -i flags will also display those packages in all repositories that depend on this package. yet, $ p=gst-plugins-bad $ diff <(pacman -Si $p) <(pacman -Sii $p) | cut -c -72 11a12,13 > Required By : cheese corebird deepin-music fractal gaupol gno > Optional For : beets clementine farstream gajim gloobus-preview 18c20,22 < Validated By : MD5 Sum SHA-256 Sum Signature --- > MD5 Sum : 33cbeb2fcb73beb293bd89525fc32139 > SHA-256 Sum : a56e743b23268133beceb9c012b93346104505b46a6b399ee385 > Signatures : A5E9288C4FA415FA $ I was searching for `option' when reading pacman(1) so would have stood a better chance of realising -Sii would help had it been there. -- Cheers, Ralph.
participants (3)
-
Andrew Gregory
-
Eli Schwartz
-
Ralph Corderoy